The Keep. Eastern Illinois University

Transcription

1 Eastern Illinois University The Keep Masters Theses Student Theses & Publications Factors Affecting the Predator-Prey Relationship Between Predaceous Diving Beetle Larvae (Dytiscus fasciventris) and Two Anuran Tadpole Species (Bufo americanus and Hyla crucifer) Todd S. Campbell Eastern Illinois University This research is a product of the graduate program in Zoology at Eastern Illinois University. Find out more about the program. Recommended Citation Campbell, Todd S., "Factors Affecting the Predator-Prey Relationship Between Predaceous Diving Beetle Larvae (Dytiscus fasciventris) and Two Anuran Tadpole Species (Bufo americanus and Hyla crucifer)" (1986). Masters Theses This Thesis is brought to you for free and open access by the Student Theses & Publications at The Keep. It has been accepted for inclusion in Masters Theses by an authorized administrator of The Keep. For more information, please contact tabruns@eiu.edu.

2 THESIS REPRODUCTION CERTIFICATE TO: Graduate Degree Candidates who have written formal theses. SUBJECT: Permission to reproduce theses. The University Library is rece1vmg a number of requests from other institutions asking permission to reproduce dissertations for inclusion in their library holdings. Although no copyright laws are involved, we feel that professional courtesy demands that permission be obtained from the author before we allow theses to be copied. Please sign one of the following statements: Booth Library of Eastern Illinois University has my permission to lend my thesis to a reputable college or university for the purpose of copying it for inclusion in that instituhon' s library or research holdings. Date Author I respectfully request Booth Library of Eastern Illinois University not allow my thesis be reproduced because - Date Author m

3 Factors Affecting The Predator-Prey Relationship Between Predaceous Diving Beetle Larvae (Dytiscus fasciventris) And Two Anuran Tadpole Species (Bufo americanus and Hyl a crucifer). (TITLE) BY Todd s. Campbell B. s. in Zoology Eastern Illinois University 1984 THESIS SUBMITIED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF M. S. in Zoology IN THE GRADUATE SCHOOL, EASTERN ILLINOIS UNIVERSITY CHARLESTON, ILLINOIS 1986 YEAR I HEREBY RECOMMEND THIS THESIS BE ACCEPTED AS FULFILLING THIS PART OF THE GRADUATE DEGREE CITED ABOVE / <;<.,.t 19; I \\ l <Ix{, 'r DEP ARlMENNiEAD

4 ABSTRACT It has previously been demonstrated that American toad (Bufo americanus) tadpoles are more vulnerable to predation by diving beetle larvae (Dytiscus fasciventris) than are spring peeper (Hyl a crucifer) tadpoles. A laboratory study was undertaken to further delineate factors that contribute to the differential vulnerability observed. Beetle larvae are more effective tadpole predators in shallow, than in deep, water and appear to prefer to adopt a "sit-and-wait" predator strategy while clinging to emergent vegetation. Depth preference experiments in the laboratory indicated that both tadpole species prefer deep areas to shallow areas irrespective of whether a predator was present or absent. Dytiscus preferred the shallow end (O - 5 cm of water) of the depth choice tank when tadpoles were absent, but preferred deeper areas when tadpoles were present. Beetle larvae prefer to eat in shallow areas, or on vegetation near the water surface. Both tadpole species appear to be more vulnerable to predation on light colored substrates, indicating that vision may be a more important prey locating mechanism in beetle larvae than previous ly thought. Bufo tadpoles move significantly more than Hyla tadpoles, which resu lts in a higher vulnerability for Bufo. Field experiments show a positive relationship between tadpole density and the number of tadpoles a beetle larva is able to capture. Field results also suggest that late

5 stage Bufo tadpoles (which move the same amount Hyla tadpoles do) are not more vulnerable to beetle larvae than are Hyla tadpoles, further delineating the relationship between the amount of movement tadpoles exhibit and the level of tadpole vulnerability to beetle larvae. These results are discussed with reference to some evolutionary theories of predator-prey systems.

6 TABLE OF CONTENTS Page Acknowledgments i Introduction 1 Materials and Methods 10 Results 22 Discussion 28 Literature Cited 40 Tables 48 Figures 56 Appendices 68

7 ACKNOWLEDGMENTS I would first like to thank Dr. Kipp C. Kruse for being my graduate advisor. His experience in this area of research was invaluable to the production of this thesis. The financial support provided and the use of his computer and other laboratory equipment were also great assets. I would also like to thank Dr. Michael Goodrich, Dr. Edward o. Moll and Dr. Kandy Baumgardner for reviewing this paper and being on my graduate committee. Much thanks goes to Ms. Patti Roberts for carrying some of the work load during the most hectic time of the research. Thanks to Lori Campbell for having the patience to endure a spouse involved in a graduate degree, and for helping with the research. Finally, I would like to thank the person most inf luential in the production of this paper, my mother, for her never ending mora l and monetary support. i

8 INTRODUCTION The complex life cyc le of anurans has been the subject of many aspects of ecological research (see Wilbur, 1980 fo r a review). Tadpole populations occur most frequently in springtime temporary or semi-permanent ponds with rapid primary production and significant time lags in the appearance of aquatic predators. They delay somatic and reproductive development in favor of growth in order to exploit the extensive primary production of ephemeral ponds (Wassersug, 1974). A wealth of literature has recently emerged addressing the conf licting roles of predation and competition in shaping tadpole community structure (see Morin, 1983). Studies by Brockleman (1969), DeBenedictus (1974), Woodward (1982), Alford and Wilbur (1985), and Wilbur and Alford (1985) have formulated a strong argument for the role that competition plays in dictating anuran larval community structure. Other studies indicate that predator-tadpole interactions lower tadpole densities enough to eliminate the effects of competition (Calef, 1973; Heyer et. al, 1975; Heyer, 1976). Cecil and Just (1979) found that populations of small ( < 20 mm) Rana catesbeiana tadpoles are controlled primarily by invertebrate predation, and not (at least as much) by abiotic factors or food availability. The outcome of competition in larval anuran communities may 1

9 depend on the density of predators present, and in the absence of these predators, the relative species abundance may be reversed in relation to species abundance when predators are present (Morin, 1983). In some cases, entire guilds may be eliminated by predators. Another hypothesis currently being advanced is that many factors, including predation, tadpole density, order of hatching, competition, and food availability are collectively responsible for the regulation of tadpole guild composition (Wilbur, 1972; Morin, 1983; Travis et al., 1985a). For example, Neill (1968) found that different survivorship curves for gulf coast toad tadpoles (Bufo valiceps) were due to a combination of insect (dytiscid) predation and tadpole competition. Likewise, larval distributions of striped chorus frog larvae (Pseudacris triseriata) on Isle Royal, Michigan were said to result from the interactions of permanence of larval habitat, size-specific predation, and density-dependent food competition (Smith, 1983). With high primary productivity, food availability is likely to be high and although niche overlap can be high, tadpoles are most likely to be regulated by predators. Heavy predation (along with dessication of the pond) would select for rapid development, constraining the predator to breeding early in the spring (Wassersug, 1975). Also, inter- and intraspecific competition can cause growth inhibition, but predatory pressures may cause selection for 2

10 rapid development and metamorphosis at small sizes in prey, and the predator would be selected for early breeding to exploit the influx of tadpoles in a pond (Wassersug, 1975). On the contrary, by laying eggs singly throughout the summer, it is suggested that barking tree frogs (Hyla gratiosa) are dispersing young throughout space and time, and that disruptive coloration and rapid growth to a less vulnerable size may increase the survival of their tadpoles (Caldwell et. al, 1980). Anurans spend a short time in metamorphosis in order to minimize the time spent with a degenerating tail and underdeveloped legs, which are liabilities on both land and water (Wassersug and Sperry, 197 7). During the egg and larval periods, aquatic breeding anurans are exposed to many different types of vertebrate and invertebrate predators. As a response to predation, some animals have developed unpalatable or toxic secretions. Many adult anurans (Dendrobatidae, Bufonidae) are toxic. It is also known that some toad eggs are unpalatable (Licht, 1968), and that Bufo tadpoles are unpalatable to fish (Voris and Bacon, 1966; Kruse and Stone, 1984), and aquatic insects (Brodie et al. 1978; Formanowicz and Brodie, 1982). Bufo larvae (black in color) presumably are conspicuous to visual predators and gregarious for thermoregulatory and anti-predatory purposes (Waldman and Adler, 1979). They are also highly density tolerant, have 3

11 an alarm reaction, and school in a polar manner (Wassersug, 1973). It is thought that schooling may be an antipredatory behavior, because upon capture, an alarm pheremone is emmitted from the injured tadpole (Pfeiffer, 1966). Sibling recognition has been shown to occur, therefore kin selection is thought to be the mechanism behind tadpole schooling behavior (see Waldman and Adler, 1979). Unlike bufonid eggs and tadpoles, Licht (1969) found that treefrog (hylid) egg masses were palatable to salamanders (Ambystoma gracile) and fish. Hyla crucifer tadpoles are also palatable and strongly preferred over Bufo spp. tadpoles by largemouth bass (Kruse and Stone, 1984), newts and salamanders (Walters, 1975), and the predatory insects Dytiscus and Lethocerus (Formanowicz and Brodie, 1982). Spring peeper tadpoles are a sandy brown color and move much less than Bufo tadpoles, making them a highly cryptic tadpole species (Kruse, unpublished data). They are not gregarious and when disturbed they tend to swim off very quickly for a short distance, then stop abruptly and settle to the pond or tank bottom. Morin (1981) found that in predator-free enclosures, Hyla crucifer tadpoles have a low survivorship relative to other tadpole species present (suggesting inferior competitive abilities), but in enclosures containing predatory newts, the species composition was reversed, suggesting that Hyla competes successfully with other tadpole species when a 4

12 predator is present, probably due to its cryptic behaviors. Predaceous diving beetles (Dytiscus fasciventris) are relatively large invertebrates that have a life span of more than a year. Both the adults and larvae can overwinter, and the adults copulate in April (Pennak, 1978). In southern Ontario, the eggs are laid in aquatic vegetation in April, and hatch in 8-13 days. The larval period (three instars) lasts about 30 days, after which it pupates in the soil at the pond edge (James, 1969). The new generation of adults emerge in June and July (Young, 196 7). Branucci (1980) found that the subfamily Dytiscinae (which includes the genus Dytiscus) is found most frequently in ponds and ditches rich in detritus and live vegetation. Also, the larvae of diving beetles tend to stay more within the vegetation than adults do (Needham and Williamson, 1907). Dytiscus as a larva has two spiracles on the last abdominal segment, and the mandibles are sickle-shaped and have a grooved canal through which digestive juices are injected in order to pre-digest food (Pennak, 1978). Dytiscids are voracious predators, and are of economic importance to fisheries (Petersen, 1960). They feed on Chironomus and Micronecta, two important food items of fish, and on the fingerlings themselves (Bisht and Das, 1979). A dytiscid has even been observed attacking and killing a young garter snake (Drummond and Wolfe, 1981). They are considered "sit-and-wait" predators, and Kruse 5

13 (1983) has shown that D. fasciventris is an optimal forager on American toad (Bufo americanus) tadpoles. Similarly, D. verticalis is an optimal forager on wood frog larvae (Rana sylvatica) and spring peeper larvae (Hyla crucifer) (Formanowicz, 1984). The role of prey motion in the ability of an aquatic predator to detect prey is well known (Ware, 1973; Wright and O'Brien, 1982). It has previously been demonstrated that Bufo americanus tadpoles are more vulnerable to predaceous diving beetle larvae (Dytiscus fasciventris) than are Hyla crucifer tadpoles because Bufo moves substantially more than Hyla does (Kruse, unpublished data). Also, Cooke (1971) showed that tadpoles treated with DDT move significantly more than normal tadpoles and are much more vulnerable to newt predators for this reason. Although many conclusions regarding the tactile and olfactory abilities of Dytiscus have been made, the visual acuity of Dytiscus has not been addressed. All past works have addressed only the olfactory component of dytiscid foraging strategies. In many cases, an aquatic invertebrate predator with opposing mandibles can capture and consume a larger prey item than can a vertebrate predator that swallows its prey whole, as in odonates vs. salamanders (Caldwell et. al 1980) and dytiscids vs. newts and salamanders (Brodie and Formanowicz 1983). Neill (1968) states that Acilius semisulcatus (Dytiscidae) is highly size-specific in its 6

14 predation on Bufo valiceps larvae, and the ability to capture and consume prey is known to depend on the relationship of the prey size (body width) to the mandible size of the beetle larva (Young, 1967). In odonates, especially Libellula, tadpole size is a very important variable, and an increase in tadpole size results in a decrease in predation rate, and density effects do not affect this relationship (Travis et. al, 1985b). Dytiscus verticalis exibits a prey size preference that is described as a correlation between it's body size and the preferred prey's body size (Brodie and Formanowicz 1983). The Dytiscus-Bufo-Hyla system is good for studying how morphologic and behavioral variation in prey species may influence its predation vulnerability. This study is an analysis of the predator-prey relationship between larval predaceous diving beetles, Dytiscus fasciventris Say (Coleoptera : Dytiscidae) and two anuran tadpole species, Bufo americanus (Bufonidae) and Hyla crucifer (Hylidae). These three species are syntopic in ephemeral and semipermenant ponds in east-central Illinois. In this study, the following experiments were designed to ask specific questions about the dytiscid-tadpole predator-prey relationship: 1. Depth preference experiment : Does Bufo, Hyla, or Dytiscus prefer a certain depth? How much do these species overlap with respect to their spatial organization? 7

15 2. Tadpole vulnerabi lity with depth experiment: Is there a depth where Bufo and Hyla larvae are more vulnerable to Dytiscus? 3. Substrate color preference experiment: Does Bufo, Hyla, or Dytiscus prefer a certain substrate color? If so, is a darker or a lighter colored substrate preferred? 4. Tadpole vulnerability with substrate color experiment: Can Dytiscus capture tadpoles more effectively on a light colored substrate or a dark colored substrate? Is the relative vulnerability of these two species similar on these two substrate colors? 5. Structure preference experiment: Does Dytiscus prefer an area that contains vertical structure? 6. Movement experiment: Kruse (unpublished data) has shown that Bufo tadpoles move 8-10 times more than do Hyla tadpoles. Does the presence of a beetle larva change the amount of movement in either species? If so, is this change enough to cause a decrease or increase in the vulnerability of that species? Also, does the presence of either species of tadpole cause a significant change in the amount of movement a beetle larva exhibits? 7. Field vulnerability experiment: Do results from field enclosures parallel results from laboratory experiments? Do effects of density on 8

16 vulnerability levels differ between Bufo and Hyla? The roles of tadpole motion, tadpole coloration, contrast between prey and background, water depth, and microhabitat structural organization are all discussed in relationship to their evolutionary implications in this predator-prey system. 9

17 MATERIALS AND METHODS General The predators used in all experiments were final instar larvae of the predaceous diving beetle Dytiscus fasciventris Say (Coleoptera: Dytiscidae). The mean volume of 12 randomly picked larvae was 0.85 ml (s.d.= 0.12). All beetle larvae were collected between 13 May, 1985 and 3 June, 1985 from a small ephemeral pond five km south of Charleston (Coles County) in east-central Illinois. They were kept in 4 liter plastic containers with 2-3 liters of water, some gravel, and a few live segments of Elodea or Naiad. Beetle larvae were fed three tadpoles each day between 8 and 12 hours before they were to be used in an experiment. The larvae were identified using taxonomic keys (Hatch, 1928; Peterson, 1960; James, 1969; Watts, 1970; Pennak, 1978) and by letting eight pupate in the lab following procedures in Formanowicz and Brodie (1981), and later identifying the newly emerged adu lts. The two anuran prey species used in this study were tadpoles of the American toad, Bufo americanus, and the spring peeper, Hyl a crucifer. These tadpoles were all collected between 13 May, 1985 and 6 June, 1985 in ephemeral and semi-permanent ponds within 10 km of Charleston, Il. They were maintained in 40 liter aquaria by segregating them according to species and date of capture, and feeding them canned spinach ad libitum. 10

18 Tadpoles were identified (Altig, 1970) and the developmental stage of each tadpole was determined using Gosner 1960 (Appendix 2). All tadpoles used were between Gesner stages 27 and 45, corresponding to the time between rear limb bud emergence and premetamorphosis. Laboratory experiments were performed between 14 May, 1985 and 7 June, 1985 during the same time each day ( ) at room temperature (22-25 C). In all experiments, tadpole size and stage, and beetle larva size was determined after each run in order to minimize stress to the animals. measurement of size in all cases. Volume was used as a Tadpole volume was determined by blot-drying all tadpoles from each run for approximately 15 seconds, placing them in a partially filled titration buret, and measuring the amount of water displaced. Dividing the displaced water volume by the number of tadpoles in the run gave an average value for tadpole size. Tadpoles were used only once, and then preserved in 10% formalin for later identification and staging. Beetle volumes were determined the same way, except they were done individually in a larger buret, and not all individuals were preserved. An acclimation period was used for all animals in all experimental runs, and water temperature was recorded at this time. Tadpoles were acclimated for between 10 and 30 minutes, and beetle larvae were acclimated for between 10 and 20 minutes, depending on the experiment (see Appendix 1 11

19 for a summary of experiments). In experiments that included a beetle larva and tadpoles, acclimation was performed by placing tadpoles in the tank and placing a single beetle larva in a small cylindrical cage, made of window screening 5 cm in diameter, in the center of the tank or quadrat. At the start of the run, the cage was gently lifted out of the water to release the beetle larva. In studies using quadrats, animals were considered to be in the quadrat containing the most anterior portion of the head. The time of tadpole capture, and carcass release, were always noted for analyzing beetle larvae intercatch intervals and feeding durations. Sampling with replacement was used in all laboratory experiments that included a predator. Five to ten "replacement tadpoles" of the same species, stage, and size were maintained during all experiments. After a tadpole was captured, a replacement tadpole was immediately placed in the same quadrat, and the carcass was removed from the tank as soon as the beetle released it. In all experimental arenas, Red flint #30 gravel (dark brown in color) and/or White dolomite #30 gravel were used as substrate. Most tests were done both with and without structure present in the water column. When structure was used, it consisted of plastic replicas of Elodea and Ludwigia. In experiments not related to depth, five cm of water was used; this depth enables the beetle larva to 12

20 respire easily. Also, since tadpoles are known to exibit phototropic responses (Wassersug, 1973), lighting for all experiments consisted of only one 40 watt incandescent lamp situated at least one meter above the apparatus in a position that presumably would not influence tadpole activity. 1. DEPTH PREFERENCE EXPERIMENT EXPERIMENTS To test for depth preferences in predators and prey, a special depth tank was constructed from plate glass (Figure 1). It was cm long, 30.5 cm wide, and 40.6 cm deep with a slanted bottom that had a slope of The tank was filled so that some gravel in quadrat 1 was exposed to the atmosphere to represent shoreline and the deep end had a maximum depth of 25.4 cm of water. The tank was divided into six quadrats with equal surface areas of 930 cm 2 A thin layer of red flint gravel was present for every experiment except Hyla runs, in which white gravel was used because Hyla tadpoles closely approximated red flint in both color and size, and were too difficult for the investigator to observe. Two 40 watt lamps were placed one meter above quadrats 2 and 5 and were the only sources of light. In a separate set of experiments, artificial plants were added to the tank to test the effects of vertical structure presence on depth choice (Figure 2). Beetles were tested for one hour, and data consisted 13

21 of the time that a beetle larva crossed each quadrat line. This resulted in the total time (in beetle-minutes, or BM) a beetle larva spent in each of the six quadrats ( 60 BM for each run). Ten tadpoles were tested for one hour, and each minute, plus or minus five seconds, the number of tadpoles in each quadrat were counted and recorded, resulting in 60 observations of 10 tadpoles for each run. Each one hour run therefore contained 600 tadpole-minutes (TM). Runs with structure (artificial plants) present were performed in the same way. To analyze the effects of predator presence on prey depth choice and effects of prey presence on predator depth choice, ten tadpoles and one beetle larva were put in the depth tank simultaneously. Tadpole and beetle positions were recorded each minute for one hour, resulting in 60 BM and 600 TM per run. The time and quadrat number of a capture or attempt at capture (a lunge with mandibles open) was noted for use in vulnerability with depth experiments. This experiment was performed both with and without structure present. The Chi-square goodness of fit test (equal expected frequencies) was used to test for any deviations from random in the distribution of animals throughout the depth tank (alpha value = 0.05). 2. THE EFFECT OF DEPTH ON TADPOLE VULNERABILITY TO BEETLE LARVAE 14

22 The effect that depth has on the vulnerability of tadpoles to predation by beetle larvae was tested in the depth tank in two different ways. In the first method, the depth tank was used as described in depth preference tests (Figures 1 and 2) to determine if there is a specific depth at which beetle larvae will capture tadpoles most frequently. Ten tadpoles and one beetle larva were tested for one hour, recording positions each minute (resulting in 600 TM) for later analysis of depth preference with a predator present. The quadrat in which a capture, hit and miss, or attempt at capture (a lunge with mandibles opened) occurred was noted. In the second method, the depth tank was divided into three equal (surface area) quadrats: shallow (1), medium (2), and deep (3), by tight fitting opaque glass dividers (Figure 3). Runs were performed with and without structure present: when structure was used it was placed in the positions shown in Figure 2. Five tadpoles of one species and one beetle larva were put in each quadrat and run for two hours. Mixed runs, consisting of three equal sized tadpoles of each species and one beetle larva were also performed. The data for each run consisted of the number of tadpoles captured in each quadrat. These data were analyzed using the Chi-square goodness of fit test (equal expected frequencies, alpha level = 0.05). 15

23 3. SUBSTRATE COLOR PREFERENCE EXPERIMENT To test for substrate color preference in beetle larvae and tadpoles, a one meter diameter plastic wading pool was set up with brown gravel and white gravel covering exactly half of the bottom, and five cm of water. Ten approximately equal size and stage tadpoles were tested for 30 minutes, and data consisted of counts of the tadpoles in each area each minute, resulting in 300 TM per run. Beetle runs lasted 30 minutes (30 BM): the time when a beetle larva crossed over the line between quadrats was recorded, resulting in the total time the beetle spent in each quadrat. The Chi-square goodness of fit test (equal expected frequencies) with the Yates correction factor was used to determine if any significant deviations from random occurred in the distribution of animals on the different substrates (alpha value = 0.05). This experiment acted as a control for testing the vulnerability of tadpoles to beetle larvae on the two different colored substrates; the percentage of time spent in each area in this test was used in deriving Chi-square expected frequencies for the tadpole vulnerability with color experiment (see next section - Exp. 4). 4. THE EFFECT OF SUBSTRATE COLOR ON TADPOLE PREDATION VULNERABILITY TO BEETLE LARVAE To test for predation vulnerability of tadpoles on 16

24 different substrate colors, a 40 liter aquarium was divided in half (two equal sized quadrats), brown gravel was on one side and white gravel on the other (Figure 4). Backgrounds of brown cork and white paper were added to the outside of the tank to match the substrate colors and the water depth was set at five cm. Ten tadpoles and one beetle larva were placed in the tank and tested for one hour during which their positions were recorded each minute, resulting in 600 TM and 60 BM in each run. When a beetle caught a tadpole or lunged at one with mandibles opened, this event was noted with respect to the quadrat in which it occurred. Data were again analyzed using the Chi-square goodness of fit test with the Yates correction factor (a lpha value = 0.05). In this analysis, however, the expected tadpole vulnerability frequencies were derived from results of the tadpole color preference tests. First, the percent of time a tadpole species spent in each area (during color preference tests) was calculated. These percentages were then used in calculating the expected number of captures that should occur in each area if background color does not affect capture frequency (Table 4). This method of data tabulation compensates for any differences in capture frequency that result from a potential preference (of tadpoles) for a given substrate color. 5. STRUCTURE PREFERENCE EXPERIMENT To test potential preferences for microhabitat spatial 17

25 organization that may occur in Dytiscus, a 1 meter diameter wading pool was set up with a brown substrate and five cm of water. Live and artificial plants were placed on one side of the pool so that exactly one half of the pool contained structure and the other did not. Each run was one hour long, in which the position of the beetle was recorded each minute, resulting in 60 BM for each run. A Chi-square goodness of fit test (equal expected frequencies) with the Yates correction factor was used to determine if beetle larvae prefer a structured or an unstructured environment (alpha value = 0. 05). 6. MOVEMENT EXPERIMENT To test for the amount of beetle larvae movement with and without prey present, and relative movement of the two tadpole species with and without a predator present, a 40 liter aquarium was divided into eight equal size quadrats (Figure 5). A thin layer of white gravel was added so the quadrat boundaries could still be seen. Beetle larvae were tested first so tadpole odors would not influence their movement, and tank water was changed after each run in which both predator and prey were present. Results of Hrbacek (1950) and Pfeiffer (1966) show that an "alarm substance" is emitted by injured Bufo tadpoles, and that this substance alters the behavior of conspecifics. Beetles were tested for 15 minutes, and data consisted of the number of quadrats the anterior portion of their 18

26 head entered. When beetle larvae were run with tadpoles, data were of the same type, but capture times and carcass release times were also recorded so runs in which a beetle larva spent mostly eating could be differentiated from runs in which a beetle larva captured no tadpoles. Tadpoles were run in groups of 10 for 15 minutes. One tadpole in each run was random ly chosen and followed for the entire time period, recording the number of quadrats its body entered with a hand-counter. If more than one tadpole occurred in the randomly chosen quadrat, a coin was flipped to choose a tadpole random ly from that quadrat. In tests with both predator and prey present, it is possible for the selected tadpole to be captured and.eaten. In these cases, the clock was stopped and a replacement tadpole of equal stage and size was put in the tank. Another tadpole (possibly the replaced one) was then random ly chosen and followed for the remainder of the 15 minute run. The time of capture and carcass release was also recorded. Analyses of tadpole data were performed in three different ways. First, differences between Bufo and Hyla movement were analyzed using an independent t-test (alpha level = 0.05, 2-tailed). Then, differences in tadpole movement (both species) with and without a predator present were also analyzed, using the same test and alpha value. Final ly, the differences in movement between early and late stage tadpoles (Bufo only) were analyzed with a one-way 19

27 ANOVA test (alpha level = 0.05). The beetle movement data were analyzed with the Kruskal-Wallis test (alpha level = 0.05) to determine if there was a difference in beetle larvae movement with and without tadpoles present. 7. TADPOLE VULNERABILITY IN FIELD ENCLOSURES In order to expose predators to prey in more natural conditions, animals were placed in field enclosures and checked every 24 hours. The enclosures were made of 2.5 x 5.0 cm lumber framing with window screening on the inside of the frame (Figure 6) Having the frame entirely on the outside of the cage walls eliminates any unnatural cover that tadpoles would have if the frame were on the inside. Window screening keeps even small aquatic organisms out, but allows algae to enter for the tadpoles to eat. These enclosures were placed in a pond south of Charleston Il. a few days prior to experimentation so that algae could establish. Three one-meter willow (Salix sp.) branches and a 5 cm diameter log approximately 0.5 meter long were placed in each of the enclosures for all of the experiments. Tadpoles were placed in the enclosures at single species densities of 50 and 20 per enclosure and mixed species densities of 20 Bufo and 20 Hyla per enclosure. Twenty-four hours later, the tadpoles were taken back to the lab, where they were counted, sized, and preserved. 20

28 Beetle larvae were used only once in field experiments. An analysis of the field data was performed using the Chi-square goodness of fit test (equal expected frequencies) with the Yates correction factor (alpha level = 0.05). 21

29 RESULTS GENERAL Predaceous diving beetle larvae (Dytiscus fasciventris), American toad tadpoles (Bufo americanus), and spring peeper tadpoles (Hyla crucifer) were found to occur syntopically in ephemeral ponds in east-central Illinois (Coles County), and beetle larvae are known to be important predators on these tadpole species (Kruse, 1983). A total of 1096 tadpoles (676 in the laboratory; 420 in the field study) and 49 beetle larvae were used in the study. Of the 620 tadpoles (310 Bufo; 310 Hyl a) exposed to predators in the laboratory, 101 were captured and-eaten. Bufo was captured significantly more (67 captures) than was Hyla (34 captures) in a combined analysis of all laboratory experiments that included predators and prey ( 1 = 8.92; 1 d.f., p << 0.05). The amount of time beetle larvae spent feeding was analyzed only for the 66 captures in which the beetle larva finished eating the prey item before the run was completed. Beetle larvae spent an average of 30.6 minutes (s.d.= 14.6; range 9-80 minutes) eating tadpoles that had a mean size of ml ( s. d. = ). The intercatch intervals (the time between when a tadpole is discarded and another one is captured, or from the begining of the run to the first capture) were analyzed to test for differences between the amount of time it takes 22

30 beetle larvae to capture the two tadpole species. Beetle larvae captured Bufo tadpoles (n=67) in an average of 11 minutes (s.d.= 8.53), and Hyla tadpoles in an average of 22 minutes (s.d.= 19.47). These results are significantly different when tested with the Mann-Whitney U test (Z = 2.32; P = 0.020; 2-tailed), suggesting that Bufo are captured significantly faster by beetle larvae than are Hyl a. 1. DEPTH PREFERENCE In laboratory studies, both Bufo and Hyl a tadpoles of similar stages (Appendix 2) significantly preferred the deep end of the depth tank (P << 0.05) whether a predator was present or not (Table 1). The presence of structure did not affect the distribution of tadpoles in the depth tank (data not shown) In six one-hour runs, beetle larvae significantly preferred the shallow end of the depth tank (O.O - 4 cm deep) when tadpoles were absent (Table 2). In 11 runs with tadpoles present, Dytiscus occurred significantly more (P < 0.05) in depths ranging from 4.3 cm to 17.1 cm deep. When beetle foraging times are analyzed separately from feeding times during 11 runs with tadpoles present, the bias of a depth choice while eating is removed. While not eating, Dytiscus was again found in the middle quadrats 2, 3, and 4 significantly more (P << 0.05) than in other quadrats. While eating Bufo, Dytiscus was found 23

31 significantly more (P << 0.05) in quadrats one and two or floating at the surface. Although Dytiscus ate Hyla exclusively (P << 0.05) in two deep quadrats, it was always at a location in aquatic vegetation at the water surface (Table 2). After all eight captures in this experiment, the beetles immediately moved with the captured tadpole to a quadrat shallower than the one in which the capture occurred. 2. VULNERABILITY WITH DEPTH Results from 10 two-hour runs suggest that tadpoles were slightly more vulnerable to beetle larvae in shallow water than in deep water, however, these results are not statistically significant to the 0.05 level for Bufo (P > 0.05), Hyla (P > 0.05), or for both species analyzed together (P < 0.10) (Table 3). Tadpole vulnerability with depth was also summarized for depth preference experiments with predators present. Tadpoles of both species were caught slightly more in the deep half of the tank where they tended to aggregate (data not shown). Six tadpoles were caught in the three deep quadrats, and two were caught in the three shal low quadrats. Of the eight tadpoles caught, six were Bufo and two were Hyla. These data were not statistically analyzed. 3. SUBSTRATE COLOR PREFERENCE During six experimental runs of 30-minutes each, Hyla 24

32 showed no significant preference for a particular substrate color (P > 0.05), whereas Bufo did significant ly prefer being in the white area (P < 0.05) (Table 4). Dytiscus significantly preferred the area with brown substrate color (P < 0.05) during eight 30-minute runs. 4. VULNERABILITY WITH COLOR In a combined analysis, both tadpole species were significantly more vulnerable to predators (P << 0.05) in the area with white substrate and background than in the darker area (Table 5). Hyla tadpoles appear to have a higher relative vulnerability on the white area than do Bufo larvae. A correction in Chi-square expected frequencies for Bufo was made because they spent significantly more time in the white area during substrate color preference tests. Despite this, toad larvae are still captured significantly more frequently (P << 0.05) in the white area. The greater total vulnerability of spring peeper tadpoles in this experiment might be attributed to the larger number of runs performed with this species (Appendix 1). 5. STRUCTURE PREFERENCE During three runs of one hour each (180 BM), Dytiscus spent significantly more time (134 BM) in the quadrat with vertical structure present than it did in the one without structure (P << 0.05). Also, during depth preference 25

33 experiments with structure, Dytiscus spent a greater percent of time on plants when it was in deeper quadrats, and spent more time feeding in plants at the surface, than it did feeding on the bottom (Table 2). 6. MOVEMENT In general, Bufo tadpoles moved significantly more than Hyl a tadpoles of equal developmental stages (t= 4.25; d.f.=13 P < 0.05; 2-tailed test) (a vs. c - Table 6). Also, earlier developmental stages of Bufo moved significantly more than did late stages (F= ; P << 0.05) (a - Table 6). The presence of a beetle larva did not appear to affect the amount of movement Bufo tadpoles exhibit (t= 1.29; d.f.=13, P > 0.05), nor does it effect the amount of movement Hyl a tadpoles exhibited (t= 0.82; d.f.= 13, P >> 0.05) (b and d - Table 6). Results from beetle movement experiments suggest that beetle larvae moved slightly less when tadpoles were present in the arena, even when the runs in which a beetle was eating a tadpole are excluded (Table 7). These results are not significant, however, using the Kruskal-Wa llis test (H = 2.93; d.f.= 2, P > 0.05). FIELD VULNERABILITY As expected, more tadpoles of both species were killed in the enclosures containing higher tadpole densities. No significant difference occurred (P > 0.05) between the 26

34 vulnerabilities of the two species in this experiment at either density, however (Table 8). The toad tadpoles were, however, at a later stage of development than were the spring peeper tadpo les (Appendix 2), and the relative movements of the two species were presumed to be approximately equal. 27

35 DISCUSSION DEPTH PREFERENCE It is known that tadpole deve lopmental time is a function of temperature, and that the relationship is a positive one, ie. the higher the temperature, the faster the development (Wilbur and Collins, 1973). For this reason, tadpoles should generally occur more frequently in warmer areas of a pond in order to speed development during this highly vulnerable period in their life cycle. Bufo is known to aggregate in kin groups and it has been found that water temperatures within these groups are significantly higher than temperatures in the surrounding water (Wassersug, 1973). From this information, it appears that tadpoles should prefer to aggregate in shallow water. My data suggest that Bufo tadpoles do not prefer shallow water, and that tadpoles do not utilize vertical structure preferentially. The reasons for these results are not entirely clear. Contrary to what is found in nature, in the depth tank, temperature differences between shallow and deep water were non-existent in any experiment, which might be a partial explanation for these results. Both species should act optimally when choosing between fast development (higher temperatures) and predator avoidance (deeper water or more cover). In the depth tank, they may have been able to sense the lack of temperature differential and were behaviorally attracted to deeper water to optimize predator 28

36 avoidance (the volume would increase, therefore decreasing tadpole density). It is also possible that anuran larvae ended up at the bottom of the gradient in their random search for warm water (anecdotal observations indicated that tadpoles were primarily negatively buoyant, and needed to swim to maintain their position). Beetle larvae have two caudal spiracles through which they must renew their air supply every few minutes (Pennak, 197 8), and it appeared that they used them in buoyancy compensation as well as respiration. They could be positively or negatively buoyant by varying the amount of air they take in with their two caudal spiracles. Beetle larvae sometimes stayed at the surface for long periods, but upon decent they expelled some air (which should decrease diving time) or tried to counteract positive buoyancy by constantly swimming downwards until an underwater structure was grasped and should decrease dive time). (which uses more oxygen In a sit-wait predator, the act of returning to the surface for air can be viewed as a risk (Cook and Streams, 19 84) that Dytiscus should tend to minimize, since it directly interferes with this foraging strategy. Also, it may be exposing itself to larger visual or tactile oriented predators such as ranid frogs, or fishing spiders of the genus Dolomedes, both of which were found in the study area. By remaining in shallow water or on a plant near the surface, where the beetle larva can be affixed to the substrate and able to 29

37 respire and remain motionless, costs of foraging should be minimized. Results of experiments with and without prey present indicate that Dytiscus does prefer to forage, and to eat, in shallow water (or in plants at the water surface). These results also suggest that the presence of tadpoles as prey items does affect the behavior of Dytiscus with respect to the depth at which it prefers to forage. VULNERABILITY WITH DEPTH Anecdotal observations during depth preference experiments with beetle predators present suggest that tadpole capture s occur slightly more frequently in the deep end of the tank. This may have been due to the significantly increased presence of tadpoles in that area of the tank. Consequently the depth tank was separated physically into three distinct quadrats for further vulnerability experiments. The results of these tests, though not statistically significant, indicate that beetle larvae are able to capture tadpoles slightly better in shallow areas. Combined with the decreased costs of staying in shallow water is the increased chance that a tadpole will have to swim directly past a waiting beetle, not above or below it. For these reasons, it seems logical to assume that beetle larvae may be more efficient predators on tadpoles in shallow water. The effects of using white gravel during the Hyl a runs is not known, but this procedure may have 30

38 overestimated true Hyl a vulnerability. More experiments need to be designed to further test aspects of tadpole vulnerability to Dytiscus over a range of depths. SUBSTRATE COLOR PREFERENCE This experiment was designed as a control for further vulnerability tests, but the results are interesting. The preference for the white area by Bufo tadpoles may be explained by theories of aposematism. Bufo are distasteful to vertebrate predators (Brodie et al. 1978) and to predators that must come in contact with the skin of this tadpole species. It is also jet black, swims almost constantly, and tends to aggregate with siblings in a manner consistent with kin selection theory (Wassersug, 1973). These characteristics make Bufo tadpoles a highly conspicuous animal, and theories of aposematism do not neccessarily contradict the idea that Bufo tadpoles are black to derive a thermal advantage. If kin selection is operating on tadpoles, the learning experiences of predators are of great importance to the fitness of each individual in an aggregate, and any tadpole behavior that can enhance that learning experience should be selected for over evolutionary time. One way to enhance the learning experience may be to increase the contrast between prey and background (Cook and Streams, 1984). Bufo could do this if it is able to differentiate between high and low contrast areas. My data suggest that Bufo is assessing background 31

39 color; however, this study does not concentrate on the mechanisms that are involved in this possible ability to differentiate substrate colors. As a predator on Bufo larvae, Dytiscus is at an extreme advantage over animals that swallow tadpoles whole. By piercing through the skin, it is avoiding the primary anti-predatory defense that Bufo tadpoles have. It is very likely that bufonids have evolved toxins in their skin as an anti-predator mechanism against engulfing-type predators (e.g. fish and salamanders), suggesting that predaceous beetle larvae are the primary predators on toad larvae. It is at best only speculation that the evolution of dytiscid mouthparts was in part dictated by the need to pierce through toxic tadpole skin. Larval dytiscid mouthparts are probably primarily adapted to pierce through arthropod exoskeleton and, as a result, avoiding Bufo skin secretions. could be well pre-adapted for It is also ingesting less of each prey item than a swallowing predator would (if it could handle the toxin) and must be able to capture a relatively greater number of prey items than a comparable sized swallowing predator per unit of time. Because of these advantages, Dytiscus larvae are very effective predators on Bufo tadpoles where these species are syntopic. The random distribution of Hyl a tadpoles on the two backgrounds might be explained in terms of cryptic behaviors. Hyl a is crypticly colored and palatable to 32

40 vertebrate predators that swallow their prey (Formanowicz and Brodie 1982). Natural selection theory predicts that cryptic animals should have an ability to select substrates with which to blend. Hyl a, in this experiment, did not prefer the area that would give it the most protection. It is known that Hyl a crucifer tadpoles move very little (Kruse, unpublished data) which may be considered a cryptic behavior. Remaining motionless on a substrate may therefore have been more important in the evolution of Hyla crucifer behavior than was background matching. Predaceous diving beetle larvae need to remain concealed from their own predators, and since these are probably largely sit-and-wait predators, they mu st remain concealed from their own prey. The fitness of a beetle larva would be enhanced if it were able to detect a suitable background and remain relatively motionless upon it. The results of this experiment indicate that Dytiscus is able to choose a substrate that it can blend with. Again, the proximate mechanisms have yet to be discovered (it is likely visual) by which Dytiscus chooses a spec ific substrate color. TADPOLE VULNERABILITY WITH SUBSTRATE COLOR Results of this experiment shed some light on the previous section. In past studies, Dytiscus has been referred to as a largely olfactory predator (Formanowicz, 1984). This is suggested here by beetle larvae movement 33

41 experiments, in which beetles move slightly (but not significantly) less during runs in which tadpoles are present (see tadpole movement section ). This would be expected in a sit-wait predator; when prey are absent, it should actively search more than when prey (or the scents of prey) are present (Formanowicz, 1984). The visual acuity of this species has been overlooked in past studies, however. Dytiscus larvae have six pairs of ocelli located laterally and posterior to the mandibles (Pennak, 1978), but the extent of their visual field is not known. Results herein indicate that Dytiscus ha s the ability to use vision to capture tadpoles. Firstly, during experiments of vulnerability with depth in the partitioned depth tank, both tactile and olfactory cues were controlled because the tadpoles were on the other side of a glass plate. Anecdotal observations suggest that beetle larvae interacted (or tried to) with the tadpoles on the other side of the glass. The tracking ability of the beetle larvae appeared to be quite good, and the strikes at each tadpole (through the glass) were accurate. After this was observed, the glass was taped in order to block the beetle 's view into the next quadrat. Secondly, no significant differences should occur between the brown area and the white area in the interactions of beetle larvae and tadpoles if these animals are not using visual cues to capture prey. If they are able to use vision in capturing prey, there should be a 34

42 significant increase in the number of interactions with tadpoles in an area where the contrast between prey and substrate is high (Cook and Streams, 1984). Thirdly, Hyl a did exhibit a much higher difference between vulnerability on the two areas than did Bufo, because the former are almost invisible on the brown area, wheras Bufo can be seen well on both substrate colors. Clearly, in a pond with low visibility, Dytiscus would rely on chemical/ tactile cues to locate an area where prey occur, then tactile information when actually capturing prey. In a pond that has clear water, Dytiscus might be able to use chemical cues first, and then a combination of visual and tactile information when making the capture. This conjecture is largely congruent with the results of Vinyard and O'Brien (1976), who found that turbidity in a pond decreases the visual reactive distance of fish predators, causing them to use a tactile detection method (the lateral line). The pond from which these animals were collected has quite clear water from the time ice clears until mid-summer, suggesting that Dytiscus may be using vision to capture prey in it's natural environment. MOVEMENT EXPERIMENTS In a previous study, Bufo americanus tadpoles were found to move eight to twelve times more than Hyla crucifer tadpoles (Kruse, unpublished) of equal developmental stage. This was determined to be the major reason why Bufo are 35

43 more vulnerable to beetle larvae than are Hyla. The amount of movement observed in Hyl a is approximately equal to the amount of movement observed in late developmental stage Bufo tadpole s. Hyl a appears to move very little at all developmental stages, but this was only an anecdotal observation and cannot be statistically tested from data in this study. Al so, Hyl a tadpoles are larger than equalstaged Bufo tadpoles and have a relatively longer tail. When disturbed, Hyl a tadpoles dart away at high speed, then rapidly stop swimming and sink to the bottom, where they remain motionless until disturbed again. In contrast, Bufo tadpoles appear to be in almost constant motion, which should increase its relative vulnerability substantially. These results are similar to those of Wright and O'Brien l I I (1982), who found that prey motion more than doubled the distance at which midge fly larvae (Chaoborus) could be located by white crappie predators. Results of this study show a decrease of movement in Bufo tadpoles as development proceeds. A logical explanation is that limb development tends to increase the drag on a swimming tadpole. This, coupled with the loss of tail mass makes swimming increasingly difficult, and the tadpole must necessarily slow down. In contrast, Hyl a appears to move very little irrespective of developmental stage (anecdotal observations). These findings indicate that the vulnerability of Bufo to beetle larvae may change (decrease) over time, and that Hyl a vulnerability may not. 36

44 It is therefore logical to assume that early in the tadpole stage Bufo tadpoles are much more vulnerable to beetle larvae than are Hyl a tadpoles, and that later in tadpole development (premetamorpho sis) the two species have similar vulnerabilities. A question that this study does not address is "do Bufo tadpoles move away from the alarm substance, or do they just generally move more?" A controlled experiment is needed in which a group of tadpoles are exposed to tadpole skin extract in an arena designed to reveal relative movement and tadpole locations at given points in time. Beetle larvae may (although results are not significant) tend to move more when in a tank with freshwater and no tadpoles than they should with tadpoles present (i.e., they can assess tadpole density), which is consistent with theories of optimal foraging in a sit-wait predator (Kruse, 1983). If a sit-wait predator encounters no prey, it may be beneficial to give up concealment from predators and go out foraging until it finds an area with prey (assuming prey distribution is patchy). If this is true, beetle larvae should move more in the presence of Hyla than Bufo. That Dytiscus moved slightly (but not significantly) more with Bufo present than it did in the presence of Hyl a is possibly explained by looking at the tadpole stages. In this experiment, Hyla stages are early, and Bufo stages are late (Appendix 2), making their relative movements somewhat equal, and the differences may 37

45 then be due to random variations in run s. Clearly, more experiments need to be designed to test these hypotheses further. FIELD VULNERAB ILITY This experiment was originally designed to reveal specific tadpole density effects in this predator-prey relationship. However, the only density effect observed was that beetles were able to capture and eat more tadpoles in a 24 hour period in the enclosure with higher tadpole density. Differential vulnerability results in the field are much more interesting, and relate to tadpole movement experiments. Bufo and Hyl a were equally vulnerable in this experiment, but Bufo tadpoles were well advanced in their larval development (Appendix 2), indicating that their relative movement was very similar to that of Hyla. If differential vulnerability in these tadpole species was due to factors other than the amount of movement they exhibited, this experiment should have revealed those factors by showing a difference in the two species ' vulnerjtbility. There was no difference, further suggesting that tadpole movement maybe an important determinant of tadpole vulnerability to aquatic beetle larvae. Another factor that may have been important in these results was the high turbidity of the water in the test pond. The turbid water would have decreased the reactive distance of 38

46 Dytiscus substantially, possibly forcing it to use nonvisual methods of prey location. SUMMARY In conclusion, it has been shown in this study that Bufo americanus tadpoles are generally more vulnerable to predaceous diving beetle larvae than are Hyl a crucifer tadpole s, and that this is related to the amount of movement tadpoles exhibit. The roles of tadpole movement, microhabitat utilization, and species composition of a pond over time are important in the evolution of anuran-dytiscid predator-prey interactions and deserve further study. 39

47 LITERATURE CITED Alford, R. A. and H. M. Wilbur Priority effects in experimental pond communities : Competition between Bufo and Rana. Ecology 66: Altig, R A key to the tadpoles of the continental United States and Canada. Herpetologica 26: Bisht, R. S. and S. M. Da s Observations on the food and feeding of some Dytiscidae (Coleoptera) of Kumaon lakes with notes on their ecology. J. Inland Fish. Soc. India 11: Branucci, M Observations on the ecology of the Dytiscidae in a water system annected to the south coast of the Neuchatels ' lake (Coleoptera, l I I Oytiscidae). Bul l. Soc. Entom. Suisse 53: Brockleman, W. Y An analysis of density effects and predation in Bufo americanus tadpoles. Ecology 50: Brodie, E. D., Jr. and D. R. Formanowicz, Jr Prey size preference of predators : Differential vulnerability of larval anurans. Herpetologica 39 :

48 Brodie, E. D., Jr., D. R. Formanowicz, Jr. and E. D. Brodie, III The development of noxiousness of Bufo americanus tadpoles to aquatic insect predators. Herpetologica 34: Caldwell, J. P., J. H. Thorp, and T. O. Jervey Predator-prey relationships among larval dragonflies, salamanders,and frogs. Oecologia 46: Calef, G. W Natural mortality of tadpoles in a population of Rana aurora. Ecology 54: Cecil, S. G. and J. J. Just Survival rate, population density and developement of a naturally occurring anuran larvae (Rana catesbiana) Copeia 1979: Cook, W. L. and F. A. Streams Fish predation on Notonecta (Hemiptera) : Relationship between prey risk and habitat utilization. Oecologia 64: Cooke, A. S Selective predation by newts on frog tadpoles treated with DDT. Nature 229: DeBenedictus, P. A Interspecific competition between tadpoles of Rana pipiens and R. sylvatica : An experimental field study. Ecol. Monogr. 44:

49 Drummond, H. and G. w. Wolfe An observation of a diving beetle larva (Insecta : Coleoptera : Dytiscidae) attacking and killing a garter snake, Thamnophis elega ns (Reptilia : Serpente s: Colubridae ). Coleopt. Bull. 35: Formanowicz, D. R., Jr Foraging tactics of an aquatic insect: Partial consumption of prey. Anim. Behav. 32: Formanowicz, D. R., Jr. and E. D. Brodie, Jr Prepupation behavior and pupation of the predaceous diving beetle Dytiscus verticalis Say (Coleoptera : Dytiscidae). N. Y. Entomo l. Soc. 89: Relative palatabi lities of members of a larval amphibian community. Copeia 1982: Gosner, K. L A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16: Hatch, M. H Studies on Dytiscidae. Brook. Entom. Soc. 23: Heyer, W. R Studies in larval amphibian habitat partitioning. Smith. Cont. Zool

50 Heyer, W. R., R. W. McDiarmid and D. L. Weigmann Tadpoles, predation and pond habitats in the tropics. Biotropica 7: Hrbacek, J On the flight reaction of tadpoles of the common toad caused by chemical substances. Experimentia 6: James, H. G Immature stages of five diving beetles (Coleoptera : Dyti scidae), notes on their habits and life history, and a key to aquatic beetles of vernal woodland pools in southern Ontario. Proc. Ent. Soc. Ont. 100: Kruse, K. C Optimal foraging by predaceous diving beetle larvae on toad tadpoles. Oecologia 58 : Kruse, K. C. and B. M. Stone Largemouth bass (Micropterus salmoides) learn to avoid feeding on toad (Bufo) tadpoles. Anim. Behav. 32: Kruse, K. C., R. J. Gawlick and P. M. Acker. In prep. The relationship of tadpole movement to predation vulnerability by aquatic beetle larvae. Unpublished manuscript. Licht, L. E Unpalatability and toxicity of toad eggs. Herpetologica 24:

51 1969. Am. Midl. Nat. Palatability of Rana and Hyl a eggs. 82: Morin, P. J Predatory Sa lamanders reverse the outcome of competition among three species of anuran tadpoles. Science 212: Predation, competition, and the composition of larval anuran guilds. Ecol. Mon. 53 : Needham, J. G. and H. V. Williamson Observations on the natural hi story of diving beetles. Am. Nat. 41: Neill, W. E Predation on Bufo valliceps tadpoles by the predaceous diving beetle Ac ilius semisulcatus. Bull. Ecol. Soc. Arn. 49:169. Pennak, R. W Fresh-water invertebrates of the United State s. John Wiley & Sons, New York. P Peterson, A Larvae of Insects: Co leoptera, Diptera, Neuroptera, Siphonaptera, Mecoptera, Trichoptera, Part II. Edwards Brothers, Inc. Ann Arbor, Michigan. Pfeiffer, W Die verbreitung der schreckreaktion bei kau lquappen und die herkunft des schreckstoffes z. Vergl. Physiol. 52:

52 Smith, D. C Factors controlling tadpole populations of the chorus frog (Pseudacris triseriata) on Isle Roya le, Michigan. Ecology 64: Travis, J., W. H. Keen and J. Juilianna. 1985a. The effects of multiple factors on viability selection in Hyl a gr atiosa tadpoles. Evolution 39: b. The role of relative body size in a predator-prey relationship between dragonfly naiads and larval anurans. Oikos 45: Vinyard, G. L. and W. J. O'Brien Effects of light and turbidity on the reactive distance of bluegi ll I! (Lepomis macrochirus ). 33: J. Fish. Re s. Board Can. Vo ris, H. K. and J. P. Bacon, Jr Differential predation on tadpoles. Copeia 1966 : Waldman, B. and K. Adler Toad tadpoles associate preferentially with siblings. Nature 282: Walters, B Studies of interspecific predation within an amphibian community. J. Herpeto l. 9: Ware, D. M Risk of epibenthic prey to predation by rainbow trout (Salmo gairdneri). J. Fish. Re s. Board Can. 30:

53 Wassersug, R. J Aspects of social behavior in anuran larvae, pp In Vial, J. L. ed., Evo lutionary Biology of the Anurans. Univ. Missouri Pre ss, Co lumbi a. Science 185: Evolution of anuran life cycles The adaptive signi ficance of the tadpole stage with comments on the maintenance of complex life cyc les in anurans. Am. Zool. 15: Wassersug, R. J. and D. G. Sperry The relationship of locomotion to differential predation on Pseudacris. J l triseriata (Anura: Hylidae). Ecology 58: Watts, C. H. S The larvae of some Dytiscidae (Coleoptera) from Delta, Manitoba. Can. Entomol. 102: Wilbur, H. M Competition, predation, and the structure of the Ambystoma-Rana sylvatica community. Ecology 53:3-21. Syst. 11: Complex life cycles. Ann. Rev. Ecol. Wilbur, H. M. and R. A. Alford Priority effects in experimental pond communities: Response of Hyla to Bufo and Rana. Ecology 66:

54 Wilbur, H. M. and J. P. Collins Ecological aspects of amph ibian metamorphosis. Science 182: Wright, D. I. and W. J. O'Brien Differential location of Chaoborus larvae and Daphnia by fish: The importance of motion and visible size. Am. Midl. Nat. 108: Woodward, B. D Tadpole competition in a desert Anuran community. Oecologia 54: Young, A. M Predation in the larvae of Dytiscus marginalis Linnaeus (Coleoptera: Dytiscidae). Pan- Pac. Entomol. 43: I I 1 47

55 Table 1. The number of tadpole minutes spent in each quadrat during depth preference experiments. Chi-square expected values are in parentheses. Quadrat (shallow) (deep) Species Total Chisquare Bufo n=90 ( 9 00) ( 90 0) <(900) ( 90 0) ( 900) ( 900) * Bufo w/beetle n=70 ( 70 0) ( 70 0) ( 70 O) ( 70 0) (700) ( 700) * Hy la * n =40 ( 4 00) ( 4 0 0) ( 4 00) (400) ( 40 0) ( 40 0) Hy la w/beetle n=40 ( 40 0) ( 40 0) ( 400) ( 40 0) ( 40 0) ( 4 0 0) * * p <<.05 in all. 48

56 Table 2. The number of beetle minutes spent in each quadrat of the depth tank during depth preference experiments. Chi-square expected values are in parenthese s. Quadrat Species (shallow) (deep) Total Chisquare Dytiscus alone n= ( 6 0) ( 6 O) ( 6 0) ( 6 0) ( 6 0) ( 6 0) Dytiscus w/bufo n= ( 7 O) ( 7 0) ( 7 0) ( 7 0) ( 7 0) (70) Dytiscus w/hyla n= ( 4 O) ( 4 O) ( 4 0) ( 4 0) ( 4 O) (40) Dytiscus w/both n=ll ( 110) (110) ( 110) ( 110) ( 110) ( 110) Total Dytiscus not eating n=ll ( 81) ( 81} ( 81) ( 8 1) ( 81} ( 81) Dytiscus eating Buf o n= (22.3) (22.3) (22.3} (22.3) (22.3} (22.3) Dytiscus eating Hyl a n= * 29 * 0 ( 6. 6} (6.6) ( 6. 6) ( 6. 6) (6.6} ( 6. 6) * Beetle was in structure at surface 100 % of the time p <.05 in all 49

57 Table 3. Frequency of tadpole captures by beetle larvae in each quadrat during vulnerability with depth experiments. Quadrat (shallow) (deep) Chi- Species Total square Bufo * Hy la * Total * N.S. (P >. 0 5) + N.S. (P <.10) 50

58 Table 4. The number of tadpole minutes spent in areas with brown and white gravel during substrate color preference experiments. Chi-square expected values are in parentheses. Substrate Color Species Brown White Total Chisquare Bufo * n=60 ( 9 0 0) ( 9 00) Hy la n=60 ( 900) (900) Dytiscus * n=8 (120) (120) * p < N.S. (P > 0 5) 51

59 Table 5. The number of interactions between beetle larvae and tadpoles in areas with brown background and white background during vulnerability with color experiments. Chi-square expected frequencies are shown in parentheses. Unequal expected frequencies for Bufo were derived from color preference experiments (see methods for explanation ). Substrate Color Species Brown White Total Chisquare Buf o * n=3 0 (17) ( 2 0) Hy la * n=6 0 ( 2 5) ( 25) Total * ( 4 2) ( 4 5) * p <<

60 Table 6. Buf o and Hyla tadpole movement with and without Dytiscus present. Species Mean Volume (ml) Median Stage Stage Range Mean Quadrats /minute s.d. a. Bufo b. Buf o + Dytiscus c. Hy la d. Hyl a + Dytiscus

61 Table 7. Beetle larvae movement with and without tadpoles pre sent, shown as quadrats per minute. Tadpoles present Tadpoles absent Buf o a Hyla b * 0.73 * x = x = x = 0.51 s.d. = 1.26 s.d.= 0.93 s.d.= 0.38 * = runs in which a beetle was eating a tadpole a = Bufo median stage 40.5, range b = Hyla median stage 35, range

62 Table 8. The number of tadpoles predated in 24-hour field enclosure experiments. The chi-square expected frequencies are in parentheses. Density Species Total Bufo 7 n= Hy la n=2 10 Total n=420 Chi * * * square (8) (39.5) (47.5) * p > (N.S.) 55

63 Figure 1. The depth tank apparatus used for depth preference experiments without structure. 56

64 DEPTH TANK V1.. SLOPE = CM _. CM b WATER LEVEL f 25.4 CM i

65 Figure 2. The depth tank set up for depth preference experiments with structure present. All dimensions are the same as in Figure 1. Fishing weights were clamped to the base of the plants to keep them in place. Four plants were placed in each quadrat in positions designated with an "x". 58

66 ::c I A. II.I Q 59

67 Figure 3. The depth tank, modified for tadpole vulnerability experiments in which the tadpoles and beetle larvae are confined to one specific depth range. 60

68 =.: z c( I- N u.-t l.o l z I- A... Q.. t- t o::: <:t: a... 61

69 Figure 4. The 40 liter (10 gal.) tank used in the tadpole vulnerability with substrate color experiment. 62

70 u 00 0 I./"'\ i_j_l _ 63

71 Figure 5. The 40 liter tank designed for movement experiments. Each of the eight quadrats have a surface area of cm 4 64

72 MOVEMENT TA NK 65

73 Figure 6. The field enclosure designed for vulnerability experiments. 66

74 u.:::r r-1 en 3: z:: o w Cl w z: a:::: - u 3: (/) 67

75 Appendix 1. A listing of all experiments included in this study, showing the apparatus used, acclimation times, run duration, and the number of runs performed. The number of runs in which structure was present is in parenthese s. Experiment Apparatus Acclim. time Run time # Runs 1. DEPTH PREFERENCE Buf o alone -- depth tank 30 min 1 hr 9 ( 2) Hy la alone depth tank 30 min 1 hr 4 ( 2) Bufo + Dytiscus depth tank 30 min 1 hr 7 ( 1) Hy la + Dytiscus depth tank 30 min 1 hr 4 ( 2) Dytiscus alone depth tank 30 min 1 hr 6 ( 1) 2. TADPOLE VULNERABILITY WITH DEPTH Buf o depth tank 30 min 2 hr 4 ( 2) Hy la depth tank 30 min 2 hr 4 ( 2) Bufo + Hyla depth tank 30 min 2 hr ( 2) 3. SUBSTRATE COLOR PREFERENCE Bufo wading pool 15 min 30 min 6 Hy la wading pool 15 min 30 min 6 Dytiscus wading pool 15 min 30 min 8 4. TADPOLE VULNERABILITY WITH SUBSTRATE COLOR Bufo two-color 15 min substrate tank Hy la two-color 15 min substrate tank 1 hr 1 hr STRUCTURE PREFERENCE Dytiscus wading pool 20 min 1 hr 3 68

76 Experiment Apparatus Acclim. time Run time # Runs 6. RELATIVE MOVEMENT Bufo alone -- movement tank 15 min 15 min 15 Hy la alone movement tank 15 min 15 min 10 Bufo + Dytiscus movement tank 15 min 15 min 10 Hyl a + Dytiscus movement tank 15 min 15 min 10 Dytiscus alone movement tank 10 min 15 min 5 Dytiscus + Bufo movement tank 15 min 15 min 5 Dytiscus + Hyl a movement tank 15 min 15 min 5 7. TADPOLE VULNERABILITY IN THE FIELD Bufo alone field enclosure 24 hr 3 Hyla alone field enclosure 24 hr 3 Mixed field enclosure 24 hr 6 69

77 Appendix 2. Gosner stages of tadpoles in all experiments. Cat. Vol, x # ml/tad. Stages Median Stage Range Bufo depth preference without structure Bl.11 3S S-38 B S S-40 B3.12 3S 3S S-39 B BS S 3S S.S B6.11 3S S-39 B S 3S S.S Total.11 (s.d.=.ols) Bufo depth preference with structure BSl S BS2.16 3S S 3S-41 Total.14 (s.d.=.018) 37.S Bufo depth preference with a beetle present (see also BS 2) BDl S 3S BD S 3S BD S S 35 3S BD BOS S BD S 35 3S 3S Total.11 (s.d.=.034)

78 Cat. Vol, x # ml/tad. Stages Median Stage Range Hy la depth preference without structure and Hy la + beetle depth preference H H Total.22 (s.d.=.049) Hyl a depth preference with structure and Hyla + beetle depth preference with structure Hl H Total.1 7 (s.d.=.014) Vulnerability with depth experiment Buf o without structure VD ll VD VD VD VD VD

79 Cat. Vol, x # ml/tad. Stages Median Stage Range Bufo with structure VD VD VD VD VD VD Hyl a without structure VD VD VD VD VD VD Hyl a with structure VD VD VD VD VD VD

80 Cat. Vol, x # ml/tad. Stages Median Stage Range Mixed runs with structure VD9 1 B: H: VD9 2 B: H: VD93 B: H: VDlOl B: H: VD 102 B: H: VD1 03 B: H: Bufo grand total.12 (s.d.=.021) Hyla grand total.15 (s.d.=.044) Bufo substrate color preference CBl CB CB CB CBS CB Total.10 (s.d.=.009) Hy la substrate color preference CHl CH CH CH CHS

81 Cat. Vol, x # ml/tad. Stages Median Stage Range CH Total.16 (s.d.=.020) Hy la vulnerability with substrate color VCl VC VC VC VC VC Total.15 (s.d.=.044) Bufo vulnerability with substrate color VC VC VC Total.13 (s.d.=.015) Bufo movement TMl TM TM Hy la movement TM TM

82 Cat. Vol, x # ml/tad. Stages Median Stage Range Bufo movement with beetle present TBl S TB Hy la movement with beetle present TB S TB S Beetle movement with Hyla present BMl S Beetle movement with Bufo present (see also TB3 and TBS) BM TBS Field experiment * Single species density = so FVll.13 Bufo : 33 remaining FV22.13 Bufo : 40 remaining FV Hyl a: 39 remaining FV Hyl a : 40 remaining Single species density = 20 FV33.16 Bufo : 18 remaining FV Hyla: 17 remaining S

83 Cat. # Vol, x ml/tad. Stages Median Stage Range Mixed Species Density = 50 (25 Bufo, 25 Hy la) FV21 B =.14 Bufo: 19 remaining H =.16 Hyla: 19 remaining FV13 B =.14 Bufo : 13 remaining H =.18 Hyl a: 18 remaining Mixed Species Density = 20 ( 10 Bufo, 10 Hy la) FV32 B =.15 Bufo : 9 remaining H =.19 Hyl a: 7 remaining FV4 1 B =.16 Bufo : 10 remaining H =.21 Hyla : 7 remaining FV4 2 B =.15 Bufo : 8 remaining H =.20 Hyla: 10 remaining FV4 3 B =.13 Bufo : 8 remaining H =.21 Hyla: 10 remaining Total Bufo.14 (s.d.= 0. 01) Total Hy la.19 (s.d.= 0.02) * individual stages are not shown ; remaining were staged all the tadpoles 76